JP7579536B2 - Information processing device, information processing method, and information processing program - Google Patents

Description

The present disclosure relates to an information processing device, an information processing method, and an information processing program that support learning.

In recent years, services that deliver courses for qualification exams and the like via the Internet have become widespread. Learners take courses within a limited time with the aim of passing the qualification exam.
Technologies to maintain the learner's motivation to study are being developed. If the learner is shown the test score that is predicted to be obtained at the current time if the learner were to take the test during the study period, this may lead to the learner's motivation to study being maintained.

Patent document 1 describes a method for calculating points to be awarded to or taken away from learners in order to efficiently support their learning, based on the learning behavior, test results, points, and pass/fail information of multiple learners.

Patent No. 6815673

However, the technology disclosed in Patent Document 1 does not present the learner with the predicted test score that the learner can achieve on the test.

The present invention therefore aims to provide a technology that obtains the test score that a learner is predicted to achieve on a test.

The information processing device according to the present disclosure includes a unit score processing unit that acquires a score for an output lesson associated with a unit based on the attendance of the output lesson by a target user, and updates a unit score indicating the target user's ability in the unit based on the score of the output lesson, and a test score prediction unit that acquires a predicted test score for the target user based on the unit scores of a plurality of units for the target user.

The information processing device, information processing method, and information processing program disclosed herein are intended to provide a technology for obtaining the test score that a learner is predicted to achieve on an exam.

FIG. 1 is a block diagram showing an example of the configuration of a server according to the first embodiment. FIG. 2 is a diagram showing an example of the configuration of point allocation ratio data stored in the server according to the first embodiment. FIG. 3 is a flowchart showing an example of a process based on obtaining a score of an output lesson executed by the server according to the first embodiment. FIG. 4 is a flowchart illustrating an example of a unit score update process executed by the server according to the first embodiment. FIG. 5 is a flowchart illustrating an example of a process based on the passage of time, which is executed by the server according to the first embodiment. FIG. 6 is a flowchart showing another example of the unit score update process executed by the server according to the first embodiment. FIG. 7 is a flowchart illustrating an example of a fixation level update process executed by the server according to the first embodiment. FIG. 8 is a diagram illustrating an example of a unit score update process executed by the server according to the first embodiment. FIG. 9 is a flowchart showing an example of a process for obtaining a predicted test score executed by the server according to the first embodiment. FIG. 10 is a flowchart showing an example of an optimization process for each parameter executed by the server according to the first embodiment. FIG. 11 is a flowchart illustrating an example of a test score prediction model generation process executed by the server according to the first embodiment. FIG. 12 is a block diagram illustrating an example of the configuration of a server according to the second embodiment. FIG. 13 is a diagram showing an example of the configuration of the first similarity data stored in the server according to the second embodiment. FIG. 14 is a flowchart illustrating an example of a process of acquiring a first predicted unit score executed by the server according to the second embodiment. FIG. 15 is a flowchart illustrating an example of a generation process of a first unit score prediction model executed by the server according to the second embodiment. FIG. 16 is a block diagram illustrating an example of the configuration of a server according to the third embodiment. FIG. 17 is a diagram showing an example of the configuration of the second similarity data stored in the server according to the third embodiment. FIG. 18 is a flowchart illustrating an example of a process of acquiring a second predicted unit score executed by the server according to the third embodiment. FIG. 19 is a flowchart illustrating an example of a generation process of a second unit score prediction model executed by a server according to the third embodiment. FIG. 20 is a flowchart illustrating an example of a third predicted unit score acquisition process executed by the server according to the fourth embodiment. FIG. 21 is a flowchart illustrating an example of a generation process of a third unit score prediction model executed by a server according to the fourth embodiment.

Several embodiments are described below with reference to the drawings.

[First embodiment]
(Configuration example)
FIG. 1 is a block diagram showing an example of the configuration of a server 1 according to the first embodiment.
The server 1 is an electronic device that distributes course content via a network NW to the terminals of each user who has registered to take various exam courses. The server 1 is an example of an information processing device. The network NW includes at least one of the Internet, a mobile network, and a LAN (Local Area Network), etc. Here, a user who owns a terminal 2 will be used as an example for explanation. The user who owns the terminal 2 is also referred to as a target user. The user is also referred to as a student or a learner.

The exam may be an exam in which the final result is pass/fail or an exam in which the final result is a score, and is not limited to this. For example, an exam in which the final result is pass/fail is an exam related to the granting of qualifications. An exam related to the granting of qualifications may be an exam related to the granting of national qualifications or an exam related to the granting of private qualifications. For example, an exam in which the final result is a score is an exam related to a language. The exam may be an entrance exam. Exams include tests, examinations, etc.

Each exam course is made up of multiple units for each subject that makes up the exam. A unit is a grouping of content based on a theme, etc. The lessons associated with each unit include input lessons and output lessons.

An input lesson is a lesson for a user to acquire knowledge. In one example, the input lesson is taken on the web. The input lesson may be taken by watching a video or audio. The input lesson may be taken by using an electronic text on the web. The input lesson may be taken by combining watching a video or audio and using an electronic text on the web. In this example, the server 1 can obtain the time spent by the user on the input lesson by taking a log of the user's attendance. In another example, the input lesson may be taken by using a paper text such as a booklet. In this example, the server 1 can obtain the time spent by the user on the input lesson based on the user's input.

An output lesson is a lesson for a user to confirm the retention of knowledge by solving problems. In one example, the output lesson is a format in which the user answers questions on the web. In this example, the server 1 can obtain the time spent by the user on the output lesson and the score the user obtained in the output lesson by logging the user. In another example, the output lesson is a format in which the user answers questions in an actual text such as a booklet. In this example, the server 1 can obtain the time spent by the user on the output lesson and the score the user obtained in the output lesson based on the user's input. In yet another example, the output lesson is a format in which the user answers written questions or corrected assignments. In this example, the user can submit answers to written questions or corrected assignments online by uploading document data or filling out a web form. The instructor can grade and correct the answers online and return the grading results, including the score the user obtained in the output lesson, to the user. The server 1 can obtain the score the user obtained in the output lesson based on the instructor's online grading. Server 1 can obtain the time a user takes to take an output lesson based on the user's input.

Here, the first examination for small and medium-sized enterprise management consultants will be used as an example of an exam. The first examination for small and medium-sized enterprise management consultants consists of seven subjects. The first examination for small and medium-sized enterprise management consultants is an exam with a maximum score of 700 points, with each of the seven subjects being worth 100 points. Each subject is made up of 12 units. All units of the exam that make up the course for the first examination for small and medium-sized enterprise management consultants are also called all units. All units are an example of multiple units. The course for the first examination for small and medium-sized enterprise management consultants is broadly divided into a basic ability section and a skill section. The basic ability section and the skill section include two output lessons for each unit included in each subject. The two output lessons are an output lesson on basic questions and an output lesson on past questions.

An example of the configuration of the server 1 will be described.
The server 1 includes a processor 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage device 14, and a communication interface 15. The processor 11, the ROM 12, the RAM 13, the storage device 14, and the communication interface 15 are connected to each other via a data bus or the like. In Fig. 1, the communication interface is described as "communication I/F".

Processor 11 controls the operation of the entire server 1. Processor 11 is a CPU (Central Processing Unit), but is not limited to this. Processor 11 may be configured with a circuit including a field programmable gate array, etc. Processor 11 loads a program stored in advance in ROM 12 or storage device 14 into RAM 13. Processor 11 executes the program loaded into RAM 13 to realize each part described below and execute various operations.

ROM 12 is a non-volatile memory that stores control programs, control data, etc.

RAM 13 is a volatile memory that temporarily stores data being processed by processor 11.

The storage device 14 is a device configured with a non-volatile memory that stores data, programs, etc. The storage device 14 is configured with a HDD (Hard Disk Drive) or an SSD (Solid State Drive), but is not limited to these. The storage device 14 is an example of a storage unit.

The storage device 14 stores management data 141 .
The management data 141 includes various parameters used to update the unit score and user data of each user. The unit score is a score indicating the ability of each user for each unit. For example, the unit score is a score calculated by converting the score that the user may currently obtain in the test for the range of questions in the unit into a maximum score of 100 points. The initial value indicating the unit score for each unit at the start of the course may be set based on the first ability diagnostic test of the course that indicates the performance of the target user. The initial value of the unit score for each unit may be set to a fixed value such as a minimum score.

The various parameters include difficulty, forgetting degree, contribution degree, fixation effect degree, and fixation degree.
The difficulty level is a parameter indicating the degree of difficulty of each output lesson. A higher difficulty level means that the score of the output lesson is more likely to be reflected in the unit score. This is because the user's ability tends to improve as the difficulty level increases. The score of an output lesson is the score out of 100 points that the user has acquired by taking the output lesson. The difficulty level differs for each output lesson. The difficulty level may be indicated as "A _x ". The subscript "X" of "A" is a character string that identifies the output lesson. For example, the difficulty level A _1-1a of the output lesson of the basic problem associated with unit 1 included in subject 1 is 60%. The difficulty level A _1-1b of the output lesson of the past question associated with unit 1 included in subject 1 is 80%. The difficulty level A _1-2a of the output lesson of the basic problem associated with unit 2 included in subject 1 is 60%. The difficulty level A _1-2b of the output lesson of the past question associated with unit 2 included in subject 1 is 80%. The difficulty level of the output lesson of the basic problem may not differ between units, or may differ between units. The difficulty level of the output lesson of the past question may not differ between units, or may differ between units. The difficulty levels A to _X may be changeable.

The forgetting degree is a parameter that indicates the degree to which the score decreases over time. The score decreases over time because knowledge fades from the user's memory over time. For the sake of simplicity, the forgetting degree indicates the degree to which the score decreases for each specified period. The specified period is assumed to be one day, but is not limited to this. The forgetting degree will be described as not differing between users, but may differ between users. The forgetting degree is sometimes indicated as "B". For example, the forgetting degree B is 4%. The forgetting degree B may be changeable.

The contribution degree is a parameter indicating the degree of contribution of multiple output lessons associated with all units to the unit scores of all units. A high contribution degree means that the output lesson scores are more likely to be emphasized and the current unit score is more likely to be disregarded. A high contribution degree means that the output lesson scores are more likely to be reflected in the unit scores. A low contribution degree means that even if a high score is obtained in an output lesson, it is less likely to be reflected in the unit score. Here, for the sake of simplicity, the contribution degree is assumed to be the same between units, but is not limited to this. The contribution degree may be different for each unit. If multiple output lessons are associated with a unit, the contribution degree may be different for each output lesson. The contribution degree is described as being the same between users, but may be different for each user. The contribution degree may be indicated by "C". The contribution degree C is 50%. The contribution degree C may be changeable.

The degree of retention is a parameter indicating the degree of retention of the output lesson. The degree of retention increases according to the number of times the same output lesson is taken. For example, the degree of retention increases by 30% each time the number of times the same output lesson is taken increases by one. The degree of increase in the degree of retention is not limited to this. The degree of increase in the degree of retention may vary according to the number of times the lesson is taken. If the number of times the same output lesson is taken is 0, the degree of retention is 0%. If the number of times the same output lesson is taken is 1, the degree of retention is 30%. A higher retention rate means that it is easier to suppress the decrease in score over time. A lowering of the decrease in score means that the knowledge obtained in the output lesson is hard to forget due to retention. Although the degree of retention is described as being the same between users, it may be different for each user. The degree of retention may be indicated as "D _Y ". The subscript "Y" of "D" is a character string that identifies the output lesson. The degree of retention D _Y may be changeable.

The fixation effect level is a parameter that indicates the degree to which fixation is applied. A higher fixation effect level means that it is easier to reflect the fixation effect, and therefore easier to suppress a decrease in score over time. The fixation effect level is described as being consistent between users, but it may be different for each user. The fixation effect level is sometimes indicated by "E." The fixation effect level E may be changeable.

The management data 141 includes user data for each user.
The user data includes a history of unit scores for all units. The unit score history includes an initial value. The unit score history includes a unit score for each update process performed by the unit score processing unit 111, which will be described later.
The user data includes a course attendance log. The attendance log includes the attendance date and time of each lesson that has been taken. The attendance log includes the attendance time of each lesson that has been taken. The attendance log includes the score of each output lesson based on the attendance of each output lesson. The score of the output lesson based on the attendance of the output lesson is the score of the output lesson that has been taken. If the same output lesson is taken multiple times, the attendance log includes the score of the output lesson for each attendance. If the output lesson is a mock exam, the score of the mock exam may be considered as the score of the output lesson associated with each unit and broken down. The score of the mock exam may be broken down into each unit by the question included in the mock exam. The score of a question that is linked one-to-one with each unit may be assigned to the score of that unit. The score of a question that spans multiple units may be allocated to each unit. The course attendance log is acquired by the unit score processing unit 111 based on the attendance of each lesson by the user.

The user data includes a degree of mastery _DY set for each output lesson. The degree of mastery _DY set for each output lesson varies depending on the number of times each output lesson is taken.

The user data includes predicted test scores obtained by the test score prediction unit 112 based on the unit scores of all units. The predicted test scores are test scores that indicate the total test score out of a maximum of 700 points that the user is currently predicted to be able to obtain in the test, assuming that the user takes the test. The predicted test scores may also include scores for each subject out of a maximum of 100 points. The user data may also include multiple predicted test scores obtained by the test score prediction unit 117 at different times.

The user data includes the actual test score when the user inputs the actual test score via a questionnaire form or the like via the terminal. The actual test score is a test score indicating the total test score out of a maximum of 700 points that the user actually took the test and obtained. The actual test score may include a score for each subject out of a maximum of 100 points. The actual test score may include a score for each unit. The actual test score may be a score based on self-grading or an officially notified score. The actual test score may be a test score indicating the score obtained by the user by taking a mock test conducted immediately before the actual test. In this example, the mock test score can be a substitute for or a complement to the actual test score. Collecting mock test scores is effective when the amount of data on the actual test scores is insufficient.

The user data may include user attributes if the user inputs the user attributes via the terminal. The user attributes are characteristics unique to the user. For example, the user attributes include elements such as age, gender, occupation, industry, and qualifications held.

Each user's user data may be updated based on the lessons taken or information entered by each user.

The storage device 14 stores the point allocation ratio data 142 .
The point allocation percentage data 142 is data that links each unit with the point allocation percentage allocated to each unit for each subject. The point allocation percentage is the percentage of points allocated to each unit out of the total score of 100 points for each subject. The point allocation percentage may be determined based on the number of questions given with reference to past questions, or may be determined based on the points allocated with reference to past questions. The point allocation percentage data 142 is updated in accordance with updates to the point allocation percentage.

The storage device 14 stores a test score prediction model 143 .
The test score prediction model 143 is a trained model generated by the test score prediction model processing unit 115 based on the test score prediction training data. The term "generated" includes not only a new creation mode but also an update mode. The test score prediction training data includes test score training input data and test score training output data. The test score training input data includes unit scores of multiple users for all units. The test score training input data may include the attendance time of each lesson taken by multiple users. The test score training input data may include attributes of multiple users. The test score training output data includes the actual test scores of multiple users. The test score prediction model 143 outputs test score prediction output data based on the input of the test score prediction input data. The test score prediction input data includes unit scores of the target user for all units. The test score prediction input data may include the attendance time of each lesson taken by the target user. The test score prediction input data may include attributes of the target user. The test score prediction output data includes the predicted test score of the target user. The test score learning input data is an example of learning input data or input data. The test score learning output data is an example of learning output data or output data. The test score prediction input data is an example of prediction input data or input data. The test score prediction output data is an example of prediction output data or output data.

The communication interface 15 includes various interfaces that communicatively connect the server 1 to other electronic devices using a communication protocol defined by the network NW.

The hardware configuration of server 1 is not limited to the above-mentioned configuration. Server 1 allows the omission or modification of the above-mentioned components and the addition of new components as appropriate.

Each unit realized by the processor 11 will be described.
The processor 11 realizes a unit score processing unit 111, a test score prediction unit 112, an output unit 113, an optimization processing unit 114, and a test score prediction model processing unit 115. Each unit realized by the processor 11 can also be called each function. Each unit realized by the processor 11 can also be called a control unit including the processor 11, the ROM 12, and the RAM 13.

The unit score processing unit 111 updates the unit score of the target user. The current unit score is also referred to as the unit score S _current . The unit score S _current can be subject to update processing by the unit score processing unit 111. The unit score after the update processing by the unit score processing unit 111 is also referred to as the unit score S _new .

The test score prediction unit 112 obtains a predicted test score for the target user based on the unit scores for all units for the target user. The predicted test score may include a predicted score for each subject.

The output unit 113 outputs the prediction result to the terminal 2. The prediction result includes a predicted test score.

The optimization processing unit 114 optimizes various parameters used to update unit scores.

The test score prediction model processing unit 115 generates the test score prediction model 143 through machine learning.

The terminal 2 will now be described.
The terminal 2 is an electronic device used by the target user to take the course. For example, the terminal 2 is a personal computer (PC), but may also be a smartphone or a tablet, and is not limited thereto. The terminal 2 may display the content using a web browser or an app.

The terminal 2 includes a processor 21, a ROM 22, a RAM 23, a storage device 24, a communication interface 25, an input device 26, and a display device 27. The processor 21, the ROM 22, the RAM 23, the storage device 24, the communication interface 25, the input device 26, and the display device 27 are connected to each other via a data bus or the like.

The processor 21 controls the overall operation of the terminal 2. The processor 21 is configured with the same hardware as the processor 11 described above.
The ROM 22 is configured with the same hardware as the ROM 12 described above.
The RAM 23 is configured with the same hardware as the RAM 13 described above.
The storage device 24 is configured with the same hardware as the storage device 14 described above.
The communication interface 25 is configured with the same hardware as the communication interface 15 described above.

The input device 26 is a device that accepts instructions based on operations by the target user. The input device 26 is a keyboard or a touchpad, but is not limited to these.

The display device 27 is a device that displays an image. The display device 27 is a liquid crystal display, but is not limited to this. For example, the display device 27 displays the prediction results.

The hardware configuration of terminal 2 is not limited to the above-mentioned configuration. Terminal 2 allows the omission or modification of the above-mentioned components and the addition of new components as appropriate.

FIG. 2 is a diagram showing an example of the configuration of the point allocation ratio data 142 stored in the server 1 according to the first embodiment.
For example, the point allocation ratio data 142 is a table as shown in Fig. 2. Although the table for subject 1 is shown in Fig. 2, the tables for subjects 2 to 7 are configured in the same manner as the table for subject 1.
The point allocation ratio data 142 includes a record for each unit. Each unit record includes a field for storing the point allocation ratio. "Other" is content that is not mentioned in any of the units "1-1" to "1-12" in Subject 1.

(Example of operation)
Next, an example of the operation of the server 1 configured as above will be described.
Note that the process steps described below are merely examples, and each process may be modified as much as possible. In addition, steps may be omitted, replaced, or added as appropriate for the process steps described below depending on the embodiment.

FIG. 3 is a flowchart showing an example of a process based on obtaining a score of an output lesson, which is executed by the server 1 according to the first embodiment.
The unit score processing unit 111 starts the processing shown in FIG. 3 based on the target user's taking an output lesson. The unit score processing unit 111 may start the processing immediately after taking the output lesson, or may start the processing at any timing. The output lesson taken by the target user is also called a target output lesson. The target output lesson is an example of an output lesson. A unit associated with the target output lesson is also called a target unit. The target unit is an example of a unit.

The unit score processing unit 111 obtains a score of a target output lesson based on the attendance of a target output lesson associated with a target unit by a target user (step S1). In step S1, for example, the unit score processing unit 111 obtains a score of the target output lesson based on the latest attendance of the target output lesson from the management data 141. The score of the target output lesson is also referred to as a lesson score _S1 .

The unit score processing unit 111 updates the unit score of the target unit of the target user based on the lesson score _S1 (step S2). The unit score processing unit 111 updates the unit score of the target unit by reflecting the lesson score _S1 on the unit score _Scurrent of the target unit. The unit score _Scurrent of the target unit is the unit score of the target unit immediately before taking the target output lesson.

FIG. 4 is a flowchart showing an example of a unit score update process executed by the server 1 according to the first embodiment.
FIG. 4 shows a typical example of step S2 shown in FIG.

The unit score processing unit 111 adjusts the lesson score _S1 based on the difficulty level _Ax set for the target output lesson (step S21). In step S21, for example, the unit score processing unit 111 obtains the difficulty level _Ax set for the target output lesson from the management data 141. The unit score processing unit 111 adjusts the lesson score _S1 by multiplying the lesson score _S1 by the difficulty level _Ax . The score obtained by multiplying the lesson score _S1 by the difficulty level _Ax is also referred to as the adjusted lesson score _S2 .

The unit score processing unit 111 acquires the unit score S _current of the target unit of the target user from the management data 141 (step S22).

The unit score processing unit 111 acquires the contribution degree C from the management data 141 (step S23). The contribution degree C is the contribution degree of the target output lesson to the unit score S _current of the target unit.

The unit score processing unit 111 updates the unit score of the target unit of the target user based on the adjusted lesson score _S2 and the contribution C (step S24). In step S24, for example, the unit score processing unit 111 obtains the unit score _Snew of the target unit by taking a weighted average of the adjusted lesson score _S2 and the unit score _Scurrent of the target unit based on the contribution C. For example, the calculation formula is _Snew = ( _S2 x C) + { _Scurrent x (1 - C)}. Based on the unit score _Snew , the unit score processing unit 111 updates the unit score of the target unit of the target user from the score indicated by the unit score _Scurrent to the score indicated by _Snew . The unit score processing unit 111 saves the score indicated by the unit score _Snew in the storage device 14. In addition, when the target user has a relatively high unit score and takes an output lesson with a low level of difficulty A to _X , the score indicated by S _new may become lower than the score indicated by S _current . In this case, the unit score processing unit 111 may maintain the unit score as the score indicated by the unit score S _current so that the unit score is not lowered based on taking the output lesson.

As described above, the information processing device can update the unit score of the target unit of the target user based on the lesson score _S1 of the target output lesson. This allows the information processing device to apply a mechanism that reflects the tendency of the ability of the target unit to improve by taking the target output lesson to updating the unit score of the target unit. By updating the unit score to reflect this tendency, the information processing device can obtain the test score that the user is currently predicted to be able to obtain in the test.

As described above, the information processing device can update the unit score of the target unit of the target user based on the lesson score _S2 adjusted based on the difficulty level A _X and the contribution level C. By using the difficulty level A _X and the contribution level C, the information processing device can improve the unit score of the unit by repeatedly achieving high scores in the output lessons of the same unit, rather than significantly increasing the unit score by taking one output lesson. In this way, the information processing device can apply a mechanism that reflects the tendency of the ability of the unit to gradually improve by repeatedly taking output lessons of the same unit, to the update of the unit score. By updating the unit score to reflect this tendency, the information processing device can obtain the test score that the user is currently predicted to be able to obtain in the test.

FIG. 5 is a flowchart showing an example of a process based on the passage of time, which is executed by the server 1 according to the first embodiment.
The unit score processing unit 111 starts the processing based on the passage of time shown in Fig. 5 every time a certain period of time has elapsed. The certain period of time may be the same period as the predetermined period related to the forgetting degree B. When the forgetting degree indicates the degree of daily decrease, the certain period of time may be one day. The unit score processing unit 111 may start the processing based on the passage of time at a fixed timing every day.

The unit score processing unit 111 determines whether a certain period of time has passed (step S3). In step S3, for example, the unit score processing unit 111 determines whether a certain period of time has passed since the end of the previous processing based on the passage of time. If the certain period of time has not passed (step S3, NO), the unit score processing unit 111 repeats the processing of step S3. If the certain period of time has passed (step S3, YES), the processing transitions from step S3 to step S4.

The unit score processing unit 111 updates the unit scores of all units of the target user based on the degree of score decrease over time according to the forgetting degree B (step S4). The degree of score decrease over time according to the forgetting degree B will be described later.

FIG. 6 is a flowchart showing another example of the unit score update process executed by the server 1 according to the first embodiment.
FIG. 6 shows a typical example of step S4 shown in FIG.
The unit score processing unit 111 applies the process shown in FIG. 6 to each of the unit scores of all units of the target user.

The unit score processing unit 111 acquires the unit score S _current from the management data 141 (step S41). In step S41, for example, the unit score processing unit 111 acquires the unit scores S _current in order according to the arrangement order of the units.

The unit score processing unit 111 obtains the forgetting degree B from the management data 141 (step S42).

The unit score processing unit 111 acquires the degree of fixation of the unit (step S43). The degree of fixation of the unit may be indicated as "D _Z ". The subscript "Z" of "D" is a character string that identifies the unit. The degree of fixation D _1-1 is the degree of fixation of unit 1 included in subject 1. The degree of fixation D _1-2 is the degree of fixation of unit 2 included in subject 1. In step S43, for example, the unit score processing unit 111 acquires the degree of fixation D _Y set for all output lessons associated with the unit of the acquired unit score S _current from the management data 141. The unit score processing unit 111 acquires the degree of fixation D _Z of the unit based on the degree of fixation D _Y set for all output lessons associated with the unit of the unit score S _current . The unit score processing unit 111 may acquire the degree of fixation D _Z of the unit by averaging the degree of fixation D _Y set for all output lessons associated with the _unit of the unit score S current. When only one output lesson is associated with a unit, the degree of retention _{D_Z} of the unit is equal to the degree of retention _{D_Y} set for the output lesson associated with the unit.

The unit score processing unit 111 obtains the fixation effect level E from the management data 141 (step S44).

The unit score processing unit 111 updates the unit score of the target unit of the target user based on the forgetting degree B, the unit retention degree _DZ , and the retention effect degree E (step S45). In step S45, for example, the unit score processing unit 111 acquires the unit score S _new by calculation using the calculation formula exemplified below. For example, the calculation formula is S _new = S _current × {1-B × (1-D _Z × E)}. Note that the calculation formula is not limited to this formula. B × (1-D _Z × E) is also called variable F. The variable F includes the forgetting degree B. The variable F is used in the direction of decreasing the unit score. Therefore, the variable F is an example of the degree of decrease in the score over time according to the forgetting degree B. The unit score processing unit 111 updates the unit score from the score indicated by the unit score S _current to the score indicated by S _new based on the acquisition of the unit score S _new . The unit score processing unit 111 stores the score indicated by the unit score S _new in the storage device 14. Note that the formula for updating the unit score over time does not need to include the fixation effect level E. In this example, the formula may be S _new = S _current × {1 - B × (1 - D _Z )}.

As described above, the information processing device can update the unit scores of multiple units of the target user based on the degree of score decrease over time according to the forgetting degree B. The information processing device can apply a mechanism that reflects the tendency of knowledge to be lost from the memory of the target user over time to updating the unit scores. By updating the unit scores to reflect this tendency, the information processing device can obtain the test score that the user is currently predicted to be able to obtain on the test.

FIG. 7 is a flowchart showing an example of a process for updating the fixation level _DZ executed by the server 1 according to the first embodiment.
The unit score processing unit 111 starts the process shown in Fig. 7 based on the attendance of an output lesson by a target user. The unit score processing unit 111 may start the process immediately after the attendance of an output lesson, or may start the process at any timing.

The unit score processing unit 111 identifies the target output lesson (step S5). In step S5, for example, the unit score processing unit 111 identifies the target output lesson for which the output lesson score has been stored in the storage device 14 based on the target user's attendance of the output lesson.

The unit score processing unit 111 updates the degree of retention _DY set for the target output lesson based on the attendance of the target output lesson (step S6). In step S6, for example, the unit score processing unit 111 updates the degree of retention _DY set for the target output lesson so as to increase it according to the number of attendances of the target output lesson. The unit score processing unit 111 stores the degree of retention _DY after the update process in the storage device 14 in association with the target output lesson.

The above-mentioned variable F includes the retention degree _DY . The variable F changes according to the change in the retention degree _DY . As the retention degree _DZ increases, the variable F decreases. Since the retention degree _DZ increases with each attendance of the target output lesson, the variable F decreases with each attendance of the target output lesson. The retention degree _DY acts in a direction to suppress the decrease in score over time. Updating the retention degree _DY set for the target output lesson to be increased based on the attendance of the target output lesson is an example of decreasing the rate of decrease for the target unit based on the attendance of the target output lesson.

As described above, the information processing device can reduce the degree of attrition for the target unit based on taking the target output lesson. The information processing device can apply a mechanism to update the unit score that reflects the tendency for repeated taking of output lessons for the same unit to increase retention and make necessary knowledge less likely to be forgotten. By updating the unit score to reflect this tendency, the information processing device can obtain the test score that the user is currently predicted to be able to obtain on the test.

FIG. 8 is a diagram showing an example of a unit score update process executed by the server 1 according to the first embodiment.
For simplicity of explanation, it is assumed that the target user takes one output lesson per day. Although it is assumed that the target user takes the output lessons in ascending order of lesson numbers, the order in which the output lessons are taken may be selected arbitrarily. It is also possible to take the same output lesson multiple times. For simplicity of explanation, the unit scores are shown only for Unit 1-1 and Unit 1-2 of Subject 1.

The "learning flow" indicates a basic ability section or a skill section as a major division of the learned output lessons.
"Unit" indicates a unit associated with the learned output lesson.
"Lesson" indicates a learned output lesson. The learned output lessons are arranged from top to bottom in the order in which the target user has taken them.
"Difficulty level (A _x )" indicates the difficulty level A _x of the output lesson indicated by "Lesson".

"Unit score immediately before (S _current )" indicates the unit score S _current of the unit indicated in "Unit" immediately before taking the output lesson.
"Lesson Score (S ₁ )" indicates the lesson score S ₁ based on taking the output lesson indicated by "Lesson."
"Difficulty-adjusted lesson score ( _S2 )" indicates the adjusted lesson score S2 obtained by multiplying the lesson score _S1 indicated by "Lesson score ( _S1 )" by the difficulty level _Ax indicated by "Difficulty ( _{Ax )} ."
The "weighted average unit score (S _new )" indicates the unit score S new obtained by taking the weighted average of the adjusted lesson score S ₂ indicated by the "difficulty-adjusted lesson score (S ₂ )" and the unit score S _current indicated by the "unit immediately prior score (S _{current )} ."

"Unit 1-1 Total" under "Score" shows the progress of the unit score for unit 1-1. The initial unit score for unit 1-1 is 30 points, which is set based on the ability diagnostic test. The unit score for unit 1-1 may increase based on taking the output lesson associated with unit 1-1, or it may decrease over time.

The "1-1a" in "Score" indicates the progress of the score for Output Lesson 1-1a. The initial score for Output Lesson 1-1a is 50 points, calculated by dividing the initial unit score for Unit 1-1, which is 30 points based on the ability diagnostic test, by the difficulty level A _1-1a (60%) for Output Lesson 1-1a. The score for Output Lesson 1-1a may increase based on taking Output Lesson 1-1a, or may decrease over time.

The "1-1b" in "Score" indicates the progress of the score in Output Lesson 1-1b. The initial score in Output Lesson 1-1b is 37.5 points (30 points), which is set as the initial unit score for Unit 1-1 based on the ability diagnostic test, divided by the difficulty level A _1-1b (80%) of Output Lesson 1-1b. The score in Output Lesson 1-1b may increase based on taking Output Lesson 1-1b, or may decrease over time.

"Unit 1-2 Total" in "Score" shows the progress of the unit score for unit 1-2. The initial unit score for unit 1-2 is 30 points, which is set based on the ability diagnostic test. The unit score for unit 1-2 may increase based on taking the output lesson associated with unit 1-2, or it may decrease over time.

The "1-2a" in "Score" indicates the progress of the score of Output Lesson 1-2a. The initial score of Output Lesson 1-2a is 50 points obtained by dividing 30 points, which is set as the initial unit score of Unit 1-2 based on the ability diagnostic test, by the difficulty level A _1-2a (60%) of Output Lesson 1-2a. The score of Output Lesson 1-2a may increase based on taking Output Lesson 1-2a, or may decrease over time.

The "1-2b" in "Score" indicates the progress of the score of Output Lesson 1-2b. The initial score of Output Lesson 1-2b is 37.5 points obtained by dividing 30 points, which is set as the initial unit score of Unit 1-2 based on the ability diagnostic test, by the difficulty level A _1-2b (80%) of Output Lesson 1-2b. The score of Output Lesson 1-2b may increase based on taking Output Lesson 1-2b, or may decrease over time.

"Subject 1 total" in "Score" shows the transition of the average unit scores of the multiple units that make up Subject 1. Here, for the sake of simplicity, "Subject 1 total" shows the average unit score of Unit 1-1 shown in "Unit 1-1 total" and the unit score of Unit 1-2 shown in "Unit 1-2 total".

"Unit 1-1" in "Fixation" indicates the transition of the fixation D _1-1 in unit 1-1. The fixation D _1-1 is the average value of the fixation D _1-1a shown in "Fixation""1-1a" and the fixation D _1-1b shown in "Fixation""1-1b". The initial value of the fixation D _1-1 is 0%. The fixation D _1-1 increases as either the fixation D _1-1a or the fixation D _1-1b increases.

The "1-1a" in the "degree of retention" indicates the transition of the degree of retention D _1-1a set in the output lesson 1-1a. The initial value of the degree of retention D _1-1a is 0%. The degree of retention D _1-1a increases with each attendance of the output lesson 1-1a.

The "1-1b" in the "degree of retention" indicates the transition of the degree of retention D _1-1b set in the output lesson 1-1b. The initial value of the degree of retention D _1-1b is 0%. The degree of retention D _1-1b increases each time the output lesson 1-1b is taken.

"Unit 1-2" of "Fixation" indicates the transition of the fixation D _1-2 of unit 1-2. The fixation D _1-2 is the average value of the fixation D _1-2a shown in "Fixation""1-2a" and the fixation D _1-2b shown in "Fixation""1-2b". The initial value of the fixation D _1-2 is 0%. The fixation D _1-2 increases as either the fixation D _1-2a or the fixation D _1-2b increases.

The "1-2a" in the "degree of retention" indicates the transition of the degree of retention D _1-2a set in the output lesson 1-2a. The initial value of the degree of retention D _1-2a is 0%. The degree of retention D _1-2a increases with each attendance of the output lesson 1-2a.

The "1-2b" in the "degree of retention" indicates the transition of the degree of retention D _1-2b set in the output lesson 1-2b. The initial value of the degree of retention D _1-2b is 0%. The degree of retention D _1-2b increases with each attendance of the output lesson 1-2b.

An example of the processing of the unit score processing unit 111 for unit 1-1 will be described.
It is assumed that the target user has taken the output lesson associated with Unit 1-1 four times, as shown in "Lesson". The scores shown in "Unit 1-1a" of "Score" and "Unit 1-1b" of "Score" that are circled with dashed lines are score _S1 based on taking the output lesson.

8, the update of the unit score of unit 1-1 based on the target user's attendance of output lesson 1-1a of the basic skills edition will be described. The unit score processing unit 111 acquires the unit score S _current (30 points) of unit 1-1 immediately before attending output lesson 1-1a, which is indicated by "unit score immediately before (S _current )."

The unit score processing unit 111 obtains an adjusted lesson score S 2 (48 points) indicated by the "difficulty-adjusted lesson score ( _{S 2} )" by multiplying the score S ₁ (80 points) based on taking the output lesson 1-1a indicated by the "lesson score (S ₁ )" by the difficulty level A _1-1a (60%) of the output lesson _1-1a indicated by the "difficulty level (A x )".

The unit score processing unit 111 obtains the unit score S _new (39 points) indicated by the "unit weighted average score (S _new )" by S _new = (S ₂ × contribution C) + {S _current × (1 - contribution C)}. The unit score processing unit 111 updates the unit score for unit 1-1 indicated by "Unit 1-1 total" in "Score" from the unit score S _current (30 points) to the unit score S _new (39 points).

The following describes the update of the degree of fixation D 1- _Y set for the output lesson and the degree of fixation D 1- _Z set for the unit based on the target user's attendance of the output lesson associated with unit 1-1. The unit score processing unit 111 updates the degree of fixation D _1-1a set for the output lesson 1-1a indicated by "1-1a" in the "degree of fixation" from 0% to 30% based on the target user's attendance of the basic skills version of the output lesson 1-1a. The unit score processing unit 111 updates the degree of fixation D _1-1 indicated by "unit _1-1 " in the "degree of fixation" from 0% to the average value (15%) of the degree of fixation D 1-1a (30%) indicated by " _1-1a " in the "degree of fixation" and the degree of fixation D _1-1b (0%) indicated by "1-1b" in the "degree of fixation".

Similarly, the unit score processing unit 111 updates the fixation level D _1-1b indicated by "1-1b" in the "fixation level" from 0% to 30% based on the target user's taking the basic skills output lesson 1-1b. Based on the update of the fixation level D _1-1b , the unit score processing unit 111 updates the fixation level D _1-1 indicated by "Unit 1-1" in the "fixation level" from 15% to the average value (30%) of the fixation level D _1-1a indicated by "1-1a" in the "fixation level" and the fixation level D _1-1b indicated by "1-1b" in the "fixation level".

Similarly, the unit score processing unit 111 updates the degree of fixation D _1-1a shown in "1-1a" of "fixation" from 30% to 60% based on the target user's taking the ability-level output lesson 1-1a. Based on the update of the degree of fixation D _1-1a , the unit score processing unit 111 updates the degree of fixation D _1-1 shown in "Unit 1-1" of "fixation" from 30% to the average value (45%) of the degree of fixation D _1-1a (60%) shown in "1-1a" of "fixation" and the degree of fixation D _1-1b (30%) shown in "1-1b" of "fixation".

Similarly, the unit score processing unit 111 updates the degree of fixation D _1-1b shown in "1-1b" of "fixation" from 30% to 60% based on the target user's taking the ability-level output lesson 1-1b. Based on the update of the degree of fixation D _1-1b , the unit score processing unit 111 updates the degree of fixation D _1-1 shown in "Unit 1-1" of "fixation" from 45% to the average value (60%) of the degree of fixation D _1-1a shown in "1-1a" of "fixation" and the degree of fixation D _1-1b shown in "1-1b" of "fixation".

With reference to the portion surrounded by a solid line in FIG. 8, the update of the unit score of unit 1-1 over time will be described. The unit score processing unit 111 updates the unit score of unit 1-1 from S _current (80 points) to S _new (77.8 points). S _new (77.8 points) is calculated by S _current (80 points)×{1-forgetfulness B (4%)×(1-retention D _1-1 (30%)×retention contribution E (100%))}. Similarly, the unit score processing unit 111 updates the unit score of unit 1-1 over time from S _current (77.8 points) to S _new (75.6 points). The same applies thereafter. When the degree of fixation D _1-1 is updated, the unit score processing unit 111 uses the updated degree of fixation D _1-1 .

The transition of the unit score of unit 1-1 by the processing of the unit score processing unit 111 described above will be described. The unit score of unit 1-1 starts from the initial value (30 points). The unit score of unit 1-1 increases from the initial value (30 points) to 39 points due to the lesson score S ₁ based on the target user taking the basic ability output lesson 1-1a. The unit score of unit 1-1 increases from 39 points to 46.5 points due to the lesson score S ₁ based on the target user taking the basic ability output lesson 1-1b. After that, since the target user does not take the output lesson associated with unit 1-1, the unit score of unit 1-1 gradually decreases from 46.5 points to 39.2 points over time. After that, the unit score of unit 1-1 increases from 39.2 points to 45.2 points due to the lesson score S ₁ based on the target user taking the ability output lesson 1-1a. The unit score of unit 1-1 rises from 45.2 points to 55.0 points due to lesson score S ₁ based on the target user's attendance of output lesson 1-1b of the ability level. After that, because the target user does not attend the output lesson associated with unit 1-1, the unit score of unit 1-1 gradually drops over time from 55.0 points to 53.2 points. Note that because the degree of mastery D _1-1 of unit 1-1 has risen to 60%, the unit score of unit 1-1 is less likely to decrease compared to when the degree of mastery D _1-1 is 30%.

FIG. 9 is a flowchart showing an example of a predicted test score acquisition process executed by the server 1 according to the first embodiment.
The test score prediction unit 112 acquires the unit scores S _current of all units of the target user from the management data 141 (step S7). The unit scores S _current of all units of the target user are scores according to the update process by the unit score processing unit 111. A part or all of the unit scores S _current of all units of the target user are scores that have been updated at least once according to the update process by the unit score processing unit 111.

The test score prediction unit 112 obtains a predicted test score for the target user based on the unit scores S _current of all units of the target user (step S8). In step S8, in one example, the test score prediction unit 112 obtains a predicted test score for the target user based on the unit scores S _current of all units of the target user and the point allocation ratio. In this example, the test score prediction unit 112 obtains a score for each unit by multiplying the unit score S _current by the point allocation ratio indicated by the point allocation ratio data 142. The test score prediction unit 112 sums up the scores obtained for each unit for each subject and obtains the score for each subject out of a maximum score of 100 points. The test score prediction unit 112 sums up the scores obtained for each subject and obtains the total score of the test out of a maximum score of 700 points as the predicted test score for the target user. The test score prediction unit 112 saves the predicted test score for the target user in the storage device 14.

In another example, the test score prediction unit 112 obtains a predicted test score for the target user based on the unit scores S _current of all units of the target user and the test score prediction model 143. In this example, the test score prediction unit 112 inputs test score prediction input data to the test score prediction model 143. The test score prediction input data includes the unit scores S _current of all units of the target user. The test score prediction input data may include the attendance time of each lesson taken by the target user, which is obtained by the test score prediction unit 112 from the management data 141. The test score prediction input data may include the attributes of the target user, which are obtained by the test score prediction unit 112 from the management data 141. The test score prediction unit 112 obtains test score prediction output data from the test score prediction model 143 based on the input of the test score prediction input data to the test score prediction model 143. The test score prediction output data includes the predicted test score of the target user. The test score prediction unit 112 stores the predicted test scores of the target users in the storage device 14 .

The output unit 113 outputs the prediction result to the terminal 2 (step S9). In step S9, for example, the output unit 113 outputs the prediction result to the terminal 2 in a displayable form via the network NW. The prediction result includes the predicted test score obtained by the test score prediction unit 112. The prediction result may include the unit scores S _current of all units. The display device 27 of the terminal 2 displays the prediction result.

As described above, the information processing device can obtain a predicted test score for a target user based on the unit scores and point allocation percentages for all units of the target user. By using the point allocation percentages, the information processing device can predict the predicted test score through simple processing.

As described above, the information processing device can obtain a predicted exam score for a target user based on the unit scores for all units of the target user and the exam score prediction model. By using the exam score prediction model, the information processing device can improve the prediction accuracy of the predicted exam score.

FIG. 10 is a flowchart showing an example of the optimization process of various parameters executed by the server 1 according to the first embodiment.
The optimization processing unit 114 may start the optimization process at any timing and update various parameters.

The optimization processing unit 114 obtains predicted test scores for multiple users from the management data 141 (step S10).

The optimization processing unit 114 acquires performance test scores of a plurality of users from the management data 141 (step S11).
The optimization processing unit 114 optimizes the contribution degree C, the difficulty level A _X , and the forgetting degree B of the multiple output lessons associated with all units by machine learning based on the predicted test scores of the multiple users and the actual test scores of the multiple users (step S12). The actual test scores of the multiple users may be the test scores of the actual test. The actual test scores of the multiple users may be the test scores of the mock test as described above instead of the test scores of the actual test. The actual test scores of the multiple users may be supplemented with the test scores of the mock test for users who have not input the test scores of the actual test. In step S12, for example, the optimization processing unit 114 uses the contribution degree C, the difficulty level A _X , and the forgetting degree B as explanatory variables, and the actual test score as a target variable. The optimization processing unit 114 optimizes the contribution degree C, the difficulty level A _X , and the forgetting degree B by machine learning so as to minimize the difference between the predicted test score and the actual test score of each user. The optimization processing unit 114 obtains the optimized contribution degree C, the difficulty level A _X, and the forgetting degree B. The optimization processing unit 114 stores the optimized contribution degree C, difficulty level A/ _X , and forgetting degree B in the storage device 14. The optimization processing unit 114 may optimize at least one of the contribution degree C, difficulty level A/ _X , and forgetting degree B. The optimization processing unit 114 may optimize at least one of the fixation degree D/ _Y and fixation effect degree E in addition to at least one of the contribution degree C, difficulty level A/ _X , and forgetting degree B.

As described above, the information processing device can optimize various parameters used to update unit scores. By optimizing various parameters, the information processing device can improve the prediction accuracy of predicted test scores.

FIG. 11 is a flowchart showing an example of a process for generating the test score prediction model 143 executed by the server 1 according to the first embodiment.
The test score prediction model processing unit 115 may start the generation process of the test score prediction model 143 at any timing, and may newly create the test score prediction model 143. The test score prediction model processing unit 115 may start the generation process of the test score prediction model 143 at any timing, and may update the test score prediction model 143.

The test score prediction model processing unit 115 acquires test score prediction learning data (step S13). In step S13, for example, the test score prediction model processing unit 115 acquires test score prediction learning data including test score learning input data and test score learning output data from the management data 141. The test score prediction model processing unit 115 acquires test score learning input data including unit scores S _current of multiple users for all units. The test score prediction model processing unit 115 may acquire test score learning input data including the attendance time of each lesson that has been taken by multiple users. The test score prediction model processing unit 115 may acquire test score learning input data including attributes of multiple users. The test score prediction model processing unit 115 acquires test score learning output data including the actual test scores of multiple users.

The test score prediction model processing unit 115 generates a test score prediction model 143 by machine learning based on the test score prediction learning data (step S14). In step S14, for example, the test score prediction model processing unit 115 estimates the relationship of each element included in the test score learning input data to the test score learning output data by machine learning. The test score prediction model processing unit 115 generates the test score prediction model 143 based on the estimation. The machine learning is, but is not limited to, a neural network, etc.

The user's unit score S _current reflects the user's proficiency in each test question range, and therefore may affect the test score. The test score may vary depending on the user's unit score S _current . Therefore, the user's unit score S _current may have a certain correlation with the user's actual test score.

The amount of time a user takes each lesson affects the user's level of proficiency in each test topic, and may therefore also affect the test score. The test score may vary depending on the amount of time a user takes each lesson. Therefore, there may be a certain correlation between the lessons a user has taken and the user's actual test score.

The user's attributes affect the test aptitude and therefore may affect the test score. The test score may vary depending on the user's attributes. Therefore, there may be a certain correlation between the user's attributes and the user's actual test score. Age affects memory and therefore may affect the test score. Therefore, the test score may vary depending on the user's age. Some tests may be more suitable for men and others for women. Therefore, the test score may vary depending on the user's gender. Some tests are highly relevant to each occupation and industry. Therefore, the test score may vary depending on the user's occupation and industry. Some tests have similar question formats and scopes. A user who holds a qualification in a certain exam may get a high test score in an exam similar to this exam. Therefore, the test score may vary depending on the qualification held.

The test score prediction model processing unit 115 stores the generated test score prediction model 143 in the storage device 14 (step S15).
As described above, the information processing device can generate an exam score prediction model. By using the exam score prediction model, the information processing device can improve the prediction accuracy of the predicted exam score.

Note that, in the above example, an example was described in which the link between output lessons and units is many-to-one, but the link between output lessons and units may also be many-to-many. In many-to-one, one unit is linked to multiple output lessons, and one output lesson is linked to one unit. In many-to-many, one unit is linked to multiple output lessons, and one output lesson is linked to multiple units. For example, an output lesson called "XX comprehensive questions" may be linked to multiple units such as unit A, unit B, and unit C. In the following, one output lesson linked to multiple units is also referred to as a many-to-many output lesson.

The unit score processing unit 111 may update the unit scores of each of the multiple (here, N) target units associated with the target output lesson based on the attendance of the many-to-many target output lesson. In this example, the unit score processing unit 111 obtains a lesson score _S1 of each of the multiple target units based on the attendance of the many-to-many target output lesson by the target user. The unit score processing unit 111 updates the unit scores of each of the multiple target units based on the lesson score _S1 of each of the multiple target units.

The unit score processing unit 111 can determine the lesson score _S1 of each of the multiple target units based on the degree of influence of the target output lesson on each of the multiple target units. The degree of influence is a ratio for allocating the score of the target output lesson to each target unit. The lesson score _S1 of each of the multiple target units may be a score obtained by multiplying the score of the target output lesson by the degree of influence on each target unit.

In one example, the impact can be implemented as 1 in N. In this example, the impact on each target unit is 1 in N.

In another example, when a parameter such as the inclusion percentage of each of multiple target units is set in a target output lesson, the unit score processing unit 111 can determine the degree of influence on each target unit based on the inclusion percentage of each target unit. For example, assume that an output lesson called "Comprehensive Questions on XX" is pre-set with a 50% inclusion percentage of unit A, a 30% inclusion percentage of unit B, and a 20% inclusion percentage of unit C. The unit score processing unit 111 determines the degree of influence on each of units A, B, and C based on the respective inclusion percentages of units A, B, and C. For example, the degree of influence on unit A is 50% based on the inclusion percentages.

Furthermore, each question included in the target output lesson may be linked to a target unit. This allows the degree of influence on the target unit to be set more accurately. For example, an output lesson called "Comprehensive Questions on XX" includes 10 questions. Five of the 10 questions are linked to unit A, two are linked to unit B, and three are linked to unit C. In this example, the degree of influence on each of units A, B, and C may be a percentage obtained by accumulating the degree of influence on each target unit by each question included in the target output lesson for each target unit. When the target output lesson includes 10 questions, the degree of influence on one target unit by one question is 10%. The degree of influence on unit A linked to five questions is 50%.

Note that each question may be linked to multiple (here, n) target units, rather than one. For example, question 1 is a comparison question between units A and B, and is linked to units A and B. In one example, the impact of a question linked to multiple target units on each target unit can be implemented as 1/n. The impact of question 1 on each of units A and B is 5%, which is half the impact of a single question (10%).

In another example, when a question is set with parameters such as the respective inclusion percentages of multiple target units, the unit score processing unit 111 can determine the degree of influence of a question linked to multiple target units on each target unit based on the respective inclusion percentages of the multiple target units. Assume that the inclusion percentage of unit A in question 1 is preset to 60% and the inclusion percentage of unit B to 40%. The unit score processing unit 111 may determine the degree of influence of question 1 on unit A to be 6%, which is 60% of the 10% influence of one question.

Second Embodiment
The second embodiment is an embodiment in which a unit score of a unit that a user has not yet studied (a predicted unit score for a unit to be predicted) is predicted based on the relationship between the scores of the units and the unit scores (actual unit scores) of the units that the user has studied.
In the second embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and the description thereof will be omitted. In the second embodiment, the differences from the first embodiment will be mainly described.

FIG. 12 is a block diagram showing an example of the configuration of the server 1 according to the second embodiment.
As in the first embodiment, the storage device 14 stores management data 141, point allocation ratio data 142, and a test score prediction model 143. The unit scores S _current of all units of the target user included in the management data 141 include unit scores based on the target user's actual unit scores for at least one unit and unit scores based on the target user's first predicted unit scores for a predicted target unit different from the at least one unit.

The actual unit score is a unit score for a unit whose initial value is set based on the first ability diagnostic test of the course, and is the unit score used in the process of obtaining the first predicted unit score by the unit score processing unit 111. The actual unit score may be the initial value, or may be a unit score updated from the initial value by the update process by the unit score processing unit 111 described in the first embodiment. The unit score based on the actual unit score may be a unit score updated from the actual unit score by the update process by the unit score processing unit 111 described in the first embodiment, or may be the actual unit score, depending on the timing of the process of obtaining the first predicted unit score by the unit score processing unit 111.

The unit to be predicted is a unit for which no initial value has been set based on the first ability diagnostic test of the course, and is a unit for which the unit score processing unit 111 is to obtain a first predicted unit score. In the second embodiment, a fixed value is not set as the initial value for a unit for which no initial value has been set based on the first ability diagnostic test of the course.

The first predicted unit score is a predicted unit score obtained by the unit score processing unit 111 for the unit to be predicted. The first predicted unit score may be an initial value or a unit score after the start of the course, depending on the timing of the process of obtaining the first predicted unit score by the unit score processing unit 111. The unit score based on the first predicted unit score may be a unit score updated from the first predicted unit score by the update process by the unit score processing unit 111 described in the first embodiment, or may be the first predicted unit score, depending on the timing of the process of obtaining the first predicted unit score by the unit score processing unit 111. The first predicted unit score is an example of a predicted unit score.

The term "at least one unit" includes the meaning of at least one unit out of all units. The term "at least one unit" includes the meaning of all units for which the target user has achieved an actual unit score. The user data for the target user does not include actual test scores.

The storage device 14 stores the first similarity data 144 .
The first similarity data 144 is data indicating the relationship between the scores of units based on the actual unit scores of multiple users for all units. For example, the relationship is similarity. Similarity is the degree of similarity of unit scores between units. The degree of similarity of unit scores correlates with the degree of similarity of the contents studied between units. The more similar the contents studied between units, the higher the degree of similarity of unit scores. The relationship, such as the similarity of scores between units, differs depending on the units.

The first similarity data 144 is generated by the unit score prediction model processing unit 116 based on the first unit score prediction learning data. The first unit score prediction learning data includes the actual unit scores of multiple users for all units. The multiple users in the first unit score prediction learning data may include users who have obtained actual test scores, or may include users who have not obtained actual test scores. The multiple users in the first unit score prediction learning data may or may not include the target user. The actual unit scores of each of the multiple users may be the actual unit scores of all units, or may be the actual unit scores of a portion of all units. The first similarity data 144 is an example of similarity data. The first unit score prediction learning data is an example of unit score prediction learning data.

The storage device 14 stores a first unit score prediction model 145 .
The first unit score prediction model 145 is a model defined by the similarity between the prediction target unit and each of all units. The expression "all units" in the explanation of the first unit score prediction model 145 may be read as the remaining units excluding the prediction target unit from all units. The expression "all units" in the explanation of the first unit score prediction model 145 may be read as at least one unit.

The first unit score prediction model 145 outputs the first predicted unit score of the target user for the prediction target unit by applying the actual unit score of the target user for at least one unit to the similarity between the prediction target unit and each of all units. For example, the first unit score prediction model 145 outputs the first predicted unit score of the target user based on a value obtained by multiplying the actual unit score of the target user and the similarity for each unit. The degree to which the actual unit score contributes to the prediction of the first predicted unit score varies depending on the magnitude of the similarity. The actual unit score of the target user for at least one unit may have a certain correlation with the first predicted unit score of the target user for the prediction target unit by using the relationship such as the similarity of the scores between the units.

The following describes an example in which the unit to be predicted is units 1-12. The first unit score prediction model 145 applies the target user's actual unit score for at least one unit to the similarity between units 1-12 and each of all units. The first unit score prediction model 145 outputs the target user's first predicted unit score for units 1-12 based on the value obtained by multiplying the target user's actual unit score and the similarity for each unit.

The first unit score prediction model 145 is a trained model generated by the unit score prediction model processing unit 116 based on the first similarity data 144. The first unit score prediction model 145 outputs first unit score prediction output data based on the input of first unit score prediction input data. The first unit score prediction input data includes the target user's actual unit score for at least one unit. The first unit score prediction output data includes the target user's first predicted unit score for the unit to be predicted. The first unit score prediction model 145 is an example of a unit score prediction model. The first unit score prediction input data is an example of unit score prediction input data or input data. The first unit score prediction output data is an example of unit score prediction output data or output data.

The processor 11 realizes a unit score processing unit 111, a test score prediction unit 112, an output unit 113, an optimization processing unit 114, a test score prediction model processing unit 115, and a unit score prediction model processing unit 116. Each unit realized by the processor 11 can also be referred to as each function. Each unit realized by the processor 11 can also be referred to as being realized by a control unit including the processor 11, ROM 12, and RAM 13.

The unit score processing unit 111 updates the unit score of the target user, as in the first embodiment. The processing of the unit score processing unit 111 may be similar to that of the first embodiment, and the description thereof will be omitted. The unit score processing unit 111 obtains the first predicted unit score of the target user for the unit to be predicted.

The exam score prediction unit 112, like the first embodiment, obtains a predicted exam score for the target user based on the unit scores for all units of the target user. The processing of the exam score prediction unit 112 may be similar to that of the first embodiment, and a description thereof will be omitted.

The output unit 113 outputs the prediction result to the terminal 2, as in the first embodiment. The prediction result includes a predicted test score. The processing of the output unit 113 may be the same as in the first embodiment, and a description thereof will be omitted.

The optimization processing unit 114 optimizes various parameters used to update the unit scores, as in the first embodiment. The processing of the optimization processing unit 114 may be the same as in the first embodiment, and a description thereof will be omitted.

The test score prediction model processing unit 115 generates a test score prediction model by machine learning, as in the first embodiment. The processing of the test score prediction model processing unit 115 may be the same as in the first embodiment, and a description thereof will be omitted.

The unit score prediction model processing unit 116 generates a first unit score prediction model 145 through machine learning.

FIG. 13 is a diagram showing an example of the configuration of the first similarity data 144 stored in the server 1 according to the second embodiment.
For example, the first similarity data 144 is a table as shown in FIG.
The first similarity data 144 includes a record for each unit. Each record for each unit includes a field for storing the similarity between the unit and each of all units. The similarity is a value between 0 and 1. When the similarity of the scores between units is close to 1, the contents between the units tend to be similar. In this case, when the unit score of a certain unit is high, the unit score of another unit also tends to be high for many users. On the other hand, when the similarity of the scores between units is low to 0, the contents between the units tend to be dissimilar. In this case, even if the unit score of a certain unit is high, the unit score of another unit may be high or low depending on the user. The closer the similarity of the scores between units is to 1, the higher the correlation of the unit scores between units.

(Example of operation)
Next, an example of the operation of the server 1 configured as above will be described.
Note that the process steps described below are merely examples, and each process may be modified as much as possible. In addition, steps may be omitted, replaced, or added as appropriate for the process steps described below depending on the embodiment.

FIG. 14 is a flowchart illustrating an example of a process of acquiring a first predicted unit score executed by the server 1 according to the second embodiment.
The unit score processing unit 111 may start the process of obtaining the first predicted unit score in response to an instruction input by the target user on the terminal 2. The unit score processing unit 111 may start the process of obtaining the first predicted unit score at any timing.

The unit score processing unit 111 obtains the target user's actual unit score for at least one unit from the management data 141 (step S16).

The unit score processing unit 111 acquires the first predicted unit score of the target user for the unit to be predicted based on the score relationship between the units and the acquired actual unit score of the target user (step S17). The expression "acquired actual unit score of the target user" includes the meaning of the target user's actual unit score for at least one acquired unit. In step S17, for example, the unit score processing unit 111 acquires the first predicted unit score based on the first unit score prediction model 145 and the acquired actual unit score of the target user.

The unit score processing unit 111 inputs the first unit score prediction input data to the first unit score prediction model 145. The first unit score prediction input data includes the acquired actual unit score of the target user. The unit score processing unit 111 acquires the first unit score prediction output data from the first unit score prediction model 145 based on the input of the first unit score prediction input data to the first unit score prediction model 145. The first unit score prediction output data includes the target user's first predicted unit score for the unit to be predicted. The unit score processing unit 111 stores the first predicted unit score in the storage device 14.

The unit score processing unit 111 judges whether the first predicted unit score of the target user for all prediction target units has been obtained (step S18). All prediction target units are the remaining units excluding at least one unit among all units that is associated with the actual unit score of the target user. If the unit score processing unit 111 judges that the first predicted unit score of the target user for all prediction target units has been obtained (step S18, YES), the processing ends. If the unit score processing unit 111 judges that the first predicted unit score of the target user for all prediction target units has not been obtained (step S18, NO), the processing of step S17 is repeated. The unit score processing unit 111 repeats the processing of step S17 for each prediction target unit until the first predicted unit score of the target user for all prediction target units has been obtained.

As described above, the information processing device can obtain a first predicted unit score for a target user for a unit to be predicted based on the relationship between the scores of units based on the actual unit scores of multiple users for multiple units and the actual unit score of the target user. This allows the information processing device to predict unit scores for units for which an actual unit score has not been obtained.

As described above, the information processing device can obtain the first predicted unit score of the target user for the prediction target unit based on the first unit score prediction model and the target user's actual unit score. By using the first unit score prediction model, the information processing device can improve the prediction accuracy of the first predicted unit score.

FIG. 15 is a flowchart showing an example of a generation process of the first unit score prediction model 145 executed by the server 1 according to the second embodiment.
The unit score prediction model processing unit 116 may start the generation process of the first unit score prediction model 145 at any timing, and newly create the first unit score prediction model 145. The unit score prediction model processing unit 116 may start the generation process of the first unit score prediction model 145 at any timing, and update the first unit score prediction model 145.

The unit score prediction model processing unit 116 acquires learning data for predicting the first unit score (step S19). In step S19, for example, the unit score prediction model processing unit 116 acquires learning data for predicting the first unit score from the management data 141, the learning data including the actual unit scores of multiple users for all units.

The unit score prediction model processing unit 116 generates first similarity data 144 by machine learning based on the learning data for predicting the first unit score (step S20). In step S20, for example, the unit score prediction model processing unit 116 calculates the cosine similarity of the scores between units based on the learning data for predicting the first unit score. The unit score prediction model processing unit 116 acquires the cosine similarity of the scores between units as the similarity of the scores between units. The unit score prediction model processing unit 116 generates first similarity data 144 based on the acquisition of the similarity of the scores between units. The unit score prediction model processing unit 116 stores the first similarity data 144 in the storage device 14.

The unit score prediction model processing unit 116 generates a first unit score prediction model 145 based on the first similarity data 144 (step S21). In step S21, for example, the unit score prediction model processing unit 116 generates the above-mentioned first unit score prediction model 145.

The unit score prediction model processing unit 116 stores the generated first unit score prediction model 145 in the storage device 14 (step S22).

As described above, the information processing device can generate a first unit score prediction model. By using the first unit score prediction model, the information processing device can improve the prediction accuracy of the first predicted unit score.

The unit score prediction model processing unit 116 may update the first unit score prediction model 145 based on the score included in the actual test score of the target user for the unit to be predicted and the first predicted unit score. For example, the unit score prediction model processing unit 116 updates the first unit score prediction model 145 by a machine learning method so as to reduce the difference between the score included in the actual test score for the unit to be predicted and the first predicted unit score. The unit score prediction model processing unit 116 may define a cost function using a mean squared error or the like, and determine various parameters so as to minimize this cost function.

As described above, the information processing device can update the first unit score prediction model based on the score included in the target user's actual test score for the unit to be predicted and the first predicted unit score. By updating the first unit score prediction model, the information processing device can improve the prediction accuracy of the first predicted unit score.

The second embodiment may be configured as follows.
The information processing device is equipped with a unit score processing unit that acquires an actual unit score of a target user for at least one unit, and acquires a predicted unit score of the target user for a predicted target unit different from the at least one unit based on the score relationship between units based on the actual unit scores of multiple users for multiple units and the acquired actual unit score of the target user.

[Third embodiment]
The third embodiment is an embodiment in which a unit score of a unit that a user has not yet studied (a predicted unit score for a unit to be predicted) is predicted based on the relationship between the scores of users and the unit scores (actual unit scores) of units that have been studied by multiple other users.
In the third embodiment, the same components as those in the first or second embodiment are denoted by the same reference numerals, and the description thereof will be omitted. In the third embodiment, the differences from the first embodiment will be mainly described.

(Configuration example)
FIG. 16 is a block diagram showing an example of the configuration of the server 1 according to the third embodiment.
The storage device 14 stores management data 141, point allocation ratio data 142, and test score prediction model 143, as in the first embodiment. The unit scores S _current of all units of the target user included in the management data 141 include unit scores based on the target user's actual unit scores for at least one unit and unit scores based on the target user's second predicted unit scores for a predicted target unit different from the at least one unit. The actual unit scores and the unit scores based on the actual unit scores may be the same as in the second embodiment, and a description thereof will be omitted. The predicted target unit may be the same as in the second embodiment, and a description thereof will be omitted.

The second predicted unit score is a predicted unit score obtained by the unit score processing unit 111 for the unit to be predicted. The second predicted unit score may be an initial value or a unit score after the course has started, depending on the timing of the process of obtaining the second predicted unit score by the unit score processing unit 111. The unit score based on the second predicted unit score may be a unit score updated from the second predicted unit score by the update process by the unit score processing unit 111 described in the first embodiment, or may be the second predicted unit score, depending on the timing of the process of obtaining the second predicted unit score by the unit score processing unit 111. The second predicted unit score is an example of a predicted unit score.

The term "at least one unit" includes the meaning of at least one unit out of all units. The term "at least one unit" includes the meaning of all units for which the target user has achieved an actual unit score. The user data for the target user does not include actual test scores.

The storage device 14 stores the second similarity data 146 .
The second similarity data 146 is data indicating a relationship between the scores of users based on the actual unit scores of a plurality of users for all units. For example, the relationship is a similarity. The similarity is the degree of similarity of the unit scores between users. The relationship between users, such as the similarity of the scores, differs depending on the users.

The second similarity data 146 is generated by the unit score prediction model processing unit 116 based on the second unit score prediction learning data. The second unit score prediction learning data includes the actual unit scores of multiple users for all units. The multiple users in the second unit score prediction learning data may include users who have obtained actual test scores, or may include users who have not obtained actual test scores. The multiple users in the second unit score prediction learning data include the target user and multiple other users different from the target user. The actual unit scores of each of the multiple users may be the actual unit scores of all units, or may be the actual unit scores of a portion of all units. The second similarity data 146 is an example of similarity data. The second unit score prediction learning data is an example of unit score prediction learning data.

The storage device 14 stores a second unit score prediction model 147 .
The second unit score prediction model 147 is a model defined by the similarity between the target user and each of the multiple users. The term "multiple users" in the description of the second unit score prediction model 147 may be read as multiple other users. The term "multiple users" in the description of the second unit score prediction model 147 may be read as other users who have acquired an actual unit score for the prediction target unit among the multiple other users. The second unit score prediction model 147 applies the actual unit scores of the multiple other users for the prediction target unit to the similarity between the target user and each of the multiple users, thereby outputting the target user's second predicted unit score for the prediction target unit. For example, the second unit score prediction model outputs the target user's second predicted unit score based on a value obtained by multiplying the actual unit score of the other user by the similarity for each other user. The degree to which the actual unit score contributes to the prediction of the second predicted unit score varies depending on the magnitude of the similarity. The actual unit scores of multiple other users for the unit to be predicted may have a certain correlation with the second predicted unit of the target user for the unit to be predicted by using relationships such as the similarity of scores between users.

An example will be described in which the unit to be predicted is units 1-12. The second unit score prediction model 147 applies the actual unit scores of multiple other users for units 1-12 to the similarity between the target user and each of the multiple users. The second unit score prediction model 147 outputs a second predicted unit score for the target user in units 1-12 based on a value obtained by multiplying the actual unit score and similarity of each other user.

The second unit score prediction model 147 is a trained model generated by the unit score prediction model processing unit 116 based on the second similarity data 146. The second unit score prediction model 147 outputs second unit score prediction output data based on the input of second unit score prediction input data. The second unit score prediction input data includes actual unit scores of multiple other users for the unit to be predicted. The second unit score prediction output data includes the target user's second predicted unit score for the unit to be predicted. The second unit score prediction model 147 is an example of a unit score prediction model. The second unit score prediction input data is an example of unit score prediction input data or input data. The second unit score prediction output data is an example of unit score prediction output data or output data.

The processor 11 realizes a unit score processing unit 111, a test score prediction unit 112, an output unit 113, an optimization processing unit 114, a test score prediction model processing unit 115, and a unit score prediction model processing unit 116. Each unit realized by the processor 11 can also be referred to as each function. Each unit realized by the processor 11 can also be referred to as being realized by a control unit including the processor 11, ROM 12, and RAM 13.

The unit score processing unit 111 updates the unit score of the target user, as in the first embodiment. The processing of the unit score processing unit 111 may be similar to that of the first embodiment, and the description thereof will be omitted. The unit score processing unit 111 obtains the second predicted unit score of the target user for the predicted unit.

The exam score prediction unit 112, like the first embodiment, obtains a predicted exam score for the target user based on the unit scores for all units of the target user. The processing of the exam score prediction unit 112 may be similar to that of the first embodiment, and a description thereof will be omitted.

The output unit 113 outputs the prediction result to the terminal 2, as in the first embodiment. The prediction result includes a predicted test score. The processing of the output unit 113 may be the same as in the first embodiment, and a description thereof will be omitted.

The optimization processing unit 114 optimizes various parameters used to update the unit scores, as in the first embodiment. The processing of the optimization processing unit 114 may be the same as in the first embodiment, and a description thereof will be omitted.

The test score prediction model processing unit 115 generates a test score prediction model by machine learning, as in the first embodiment. The processing of the test score prediction model processing unit 115 may be the same as in the first embodiment, and a description thereof will be omitted.

The unit score prediction model processing unit 116 generates a second unit score prediction model 147 through machine learning.

FIG. 17 is a diagram showing an example of the configuration of the second similarity data 146 stored in the server 1 according to the third embodiment.
For example, the second similarity data 146 is a table as shown in FIG.
The second similarity data 146 includes a record for each user. Each user record includes a field for storing the similarity between the user and each of the multiple users. The similarity is a value between 0 and 1. When the similarity of the scores between users is close to 1, if a certain user has a high unit score, the unit score of another user also tends to be high. On the other hand, when the similarity of the scores between users is close to 0, even if a certain user has a high unit score, the unit score of another user may be high or low. The closer the similarity of the scores between users is to 1, the higher the correlation of the unit scores between users.

(Example of operation)
Next, an example of the operation of the server 1 configured as above will be described.
Note that the process steps described below are merely examples, and each process may be modified as much as possible. In addition, steps may be omitted, replaced, or added as appropriate for the process steps described below depending on the embodiment.

FIG. 18 is a flowchart illustrating an example of a process of acquiring a second predicted unit score executed by the server 1 according to the third embodiment.
The unit score processing unit 111 may start the process of obtaining the second predicted unit score in response to an instruction input by the target user on the terminal 2. The unit score processing unit 111 may start the process of obtaining the second predicted unit score at any timing.

The unit score processing unit 111 obtains the actual unit scores of multiple other users for the unit to be predicted from the management data 141 (step S23).

The unit score processing unit 111 acquires a second predicted unit score of the target user for the unit to be predicted based on the score relationship between users and the acquired actual unit scores of the multiple other users (step S24). The expression "acquired actual unit scores of multiple other users" includes the meaning of the acquired actual unit scores of multiple other users for the unit to be predicted. In step S24, for example, the unit score processing unit 111 acquires a second predicted unit score based on the second unit score prediction model 147 and the acquired actual unit scores of the multiple other users.

The unit score processing unit 111 inputs the second unit score prediction input data to the second unit score prediction model 147. The second unit score prediction input data includes the acquired actual unit scores of multiple other users. The unit score processing unit 111 acquires the second unit score prediction output data from the second unit score prediction model 147 based on the input of the second unit score prediction input data to the second unit score prediction model 147. The second unit score prediction output data includes the second predicted unit score of the target user for the unit to be predicted. The unit score processing unit 111 stores the second predicted unit score in the storage device 14.

The unit score processing unit 111 determines whether the second predicted unit scores of the target user for all prediction target units have been obtained (step S25). If the unit score processing unit 111 determines that the second predicted unit scores of the target user for all prediction target units have been obtained (step S25, YES), the processing ends. If the unit score processing unit 111 determines that the second predicted unit scores of the target user for all prediction target units have not been obtained (step S25, NO), the processing of steps S23 and S24 is repeated. The unit score processing unit 111 repeats the processing of steps S23 and S24 for each prediction target unit until the second predicted unit scores of the target user for all prediction target units have been obtained.

As described above, the information processing device can obtain a second predicted unit score of the target user for the unit to be predicted based on the score relationships between users based on the actual unit scores of multiple users for multiple units and the actual unit scores of multiple other users. This allows the information processing device to predict unit scores for units for which no actual unit scores have been obtained.

As described above, the information processing device can obtain the second predicted unit score of the target user for the unit to be predicted based on the second unit score prediction model and the actual unit scores of multiple other users. By using the second unit score prediction model, the information processing device can improve the prediction accuracy of the second predicted unit score.

FIG. 19 is a flowchart illustrating an example of a generation process of a second unit score prediction model executed by the server 1 according to the third embodiment.
The unit score prediction model processing unit 116 may start the generation process of the second unit score prediction model 147 at any timing, and newly create the second unit score prediction model 147. The unit score prediction model processing unit 116 may start the generation process of the second unit score prediction model 147 at any timing, and update the second unit score prediction model 147.

The unit score prediction model processing unit 116 acquires learning data for predicting the second unit score (step S26). In step S26, for example, the unit score prediction model processing unit 116 acquires learning data for predicting the second unit score from the management data 141, the learning data including the actual unit scores of multiple users for all units.

The unit score prediction model processing unit 116 generates second similarity data 146 by machine learning based on the learning data for second unit score prediction (step S27). In step S27, for example, the unit score prediction model processing unit 116 calculates the cosine similarity of scores between users based on the learning data for second unit score prediction. The unit score prediction model processing unit 116 acquires the cosine similarity of scores between users as the similarity of scores between units. The unit score prediction model processing unit 116 generates second similarity data 146 based on the acquisition of the similarity of scores between users. The unit score prediction model processing unit 116 stores the second similarity data 146 in the storage device 14.

The unit score prediction model processing unit 116 generates a second unit score prediction model 147 based on the second similarity data 146 (step S28). In step S28, for example, the unit score prediction model processing unit 116 generates the above-mentioned second unit score prediction model 147.

The unit score prediction model processing unit 116 stores the generated second unit score prediction model 147 in the storage device 14 (step S29).

As described above, the information processing device can generate a second unit score prediction model. By using the second unit score prediction model, the information processing device can improve the prediction accuracy of the second predicted unit score.

The unit score prediction model processing unit 116 may update the second unit score prediction model 147 based on the score included in the actual test score of the target user for the unit to be predicted and the second predicted unit score. For example, the unit score prediction model processing unit 116 updates the second unit score prediction model 147 by a machine learning method so as to reduce the difference between the score included in the actual test score for the unit to be predicted and the second predicted unit score. The unit score prediction model processing unit 116 may define a cost function using a mean squared error or the like, and determine various parameters so as to minimize this cost function.

As described above, the information processing device can update the second unit score prediction model based on the score included in the target user's actual test score for the unit to be predicted and the second predicted unit score. By updating the second unit score prediction model, the information processing device can improve the prediction accuracy of the second predicted unit score.

The third embodiment may be configured as follows.
The information processing device is equipped with a unit score processing unit that acquires actual unit scores of multiple other users other than the target user for the unit to be predicted, and acquires a predicted unit score of the target user for the unit to be predicted based on the score relationship between users based on the actual unit scores of multiple users including the target user and the multiple other users for the multiple units, as well as the acquired actual unit scores of the multiple other users.

[Fourth embodiment]
The fourth embodiment is an embodiment in which the first predicted unit score described in the second embodiment and the second predicted unit score described in the third embodiment are blended to obtain a third predicted unit score.
In the fourth embodiment, the same components as those in the first to third embodiments are denoted by the same reference numerals, and the description thereof will be omitted. In the fourth embodiment, the differences from the first to third embodiments will be mainly described.

(Configuration example)
FIG. 20 is a block diagram showing an example of the configuration of the server 1 according to the third embodiment.
As in the first embodiment, the storage device 14 stores management data 141, point allocation ratio data 142, and a test score prediction model 143. As in the second embodiment, the storage device 14 stores first similarity data 144 and a first unit score prediction model 145. As in the third embodiment, the storage device 14 stores second similarity data 146 and a second unit score prediction model 147.

The unit scores S _current of all units of the target user included in the management data 141 include a unit score based on the actual unit score of the target user for at least one unit and a unit score based on a third predicted unit score of the target user for a predicted target unit different from the at least one unit. The actual unit score and the unit score based on the actual unit score may be the same as in the second embodiment, and a description thereof will be omitted. The predicted target unit may be the same as in the second embodiment, and a description thereof will be omitted.

The third predicted unit score is a predicted unit score obtained by the unit score processing unit 111 for the unit to be predicted. The third predicted unit score may be an initial value or a unit score after the course has started, depending on the timing of the process of obtaining the third predicted unit score by the unit score processing unit 111. The unit score based on the third predicted unit score may be a unit score updated from the third predicted unit score by the update process by the unit score processing unit 111 described in the first embodiment, or may be the third predicted unit score, depending on the timing of the process of obtaining the third predicted unit score by the unit score processing unit 111. The third predicted unit score is an example of a predicted unit score.

The term "at least one unit" includes the meaning of at least one unit out of all units. The term "at least one unit" includes the meaning of all units for which the target user has achieved an actual unit score. The user data for the target user does not include actual test scores.

The processor 11 realizes a unit score processing unit 111, a test score prediction unit 112, an output unit 113, an optimization processing unit 114, a test score prediction model processing unit 115, and a unit score prediction model processing unit 116. Each unit realized by the processor 11 can also be referred to as each function. Each unit realized by the processor 11 can also be referred to as being realized by a control unit including the processor 11, ROM 12, and RAM 13.

The unit score processing unit 111 updates the unit score of the target user, as in the first embodiment. The processing of the unit score processing unit 111 may be the same as in the first embodiment, and the description thereof will be omitted. The unit score processing unit 111 obtains the third predicted unit score of the target user for the predicted unit.

The exam score prediction unit 112 obtains a predicted exam score for the target user based on the unit scores for all units of the target user, as in the first embodiment. The processing of the exam score prediction unit 112 may be similar to that in the first embodiment, and a description thereof will be omitted.

The output unit 113 outputs the prediction result to the terminal 2, as in the first embodiment. The prediction result includes a predicted test score. The processing of the output unit 113 may be the same as in the first embodiment, and a description thereof will be omitted.

The optimization processing unit 114 optimizes various parameters used to update the unit scores, as in the first embodiment. The processing of the optimization processing unit 114 may be the same as in the first embodiment, and a description thereof will be omitted.

The test score prediction model processing unit 115 generates a test score prediction model by machine learning, as in the first embodiment. The processing of the test score prediction model processing unit 115 may be the same as in the first embodiment, and a description thereof will be omitted.

The unit score prediction model processing unit 116 generates a first unit score prediction model 145 by machine learning, as in the second embodiment. The unit score prediction model processing unit 116 generates a second unit score prediction model 147 by machine learning, as in the third embodiment. The processing of the unit score prediction model processing unit 116 may be similar to that of the second or third embodiment, and a description thereof will be omitted.

(Example of operation)
Next, an example of the operation of the server 1 configured as above will be described.
Note that the process steps described below are merely examples, and each process may be modified as much as possible. In addition, steps may be omitted, replaced, or added as appropriate for the process steps described below depending on the embodiment.

FIG. 21 is a flowchart illustrating an example of a generation process of a third unit score prediction model executed by the server 1 according to the fourth embodiment.
The unit score processing unit 111 may start the process of obtaining the third predicted unit score in response to an instruction input by the target user on the terminal 2. The unit score processing unit 111 may start the process of obtaining the third predicted unit score at any timing.

The unit score processing unit 111 obtains the target user's actual unit score for at least one unit from the management data 141 (step S30).

The unit score processing unit 111 obtains the actual unit scores of multiple other users for the unit to be predicted from the management data 141 (step S31).

The unit score processing unit 111 obtains a first predicted unit score of the target user for the unit to be predicted based on the relationship between the scores of the units and the obtained actual unit score of the target user (step S32). The processing of step S32 may be the same as the processing of step S17 described above.

The unit score processing unit 111 obtains a second predicted unit score of the target user for the unit to be predicted based on the score relationship between users and the obtained actual unit scores of the multiple other users (step S33). The processing of step S33 may be the same as the processing of step S24 described above.

The unit score processing unit 111 acquires a third predicted unit score of the target user for the prediction target unit based on the first predicted unit score and the second predicted unit score of the target user for the prediction target unit (step S34). In step S34, for example, the unit score processing unit 111 may acquire a value obtained by adding L x the first predicted unit score and (1-L) x the second predicted unit score as the third predicted unit score. L is a value greater than 0 and less than 1. The value of L can be set appropriately. The unit score processing unit 111 stores the third predicted unit score in the storage device 14.

The unit score processing unit 111 judges whether the third predicted unit score of the target user for all prediction target units has been obtained (step S35). If the unit score processing unit 111 judges that the third predicted unit score of the target user for all prediction target units has been obtained (step S35, YES), the processing ends. If the unit score processing unit 111 judges that the third predicted unit score of the target user for all prediction target units has not been obtained (step S35, NO), the processing of steps S31, S32, S33, and S34 is repeated. The unit score processing unit 111 repeats the processing of steps S31, S32, S33, and S34 for each prediction target unit until the third predicted unit score of the target user for all prediction target units has been obtained.

As described above, the information processing device can obtain a third predicted unit score based on the first predicted unit score and the second predicted unit score of the target user for the unit to be predicted. This allows the information processing device to improve the prediction accuracy of the unit score by using two different methods.

The fourth embodiment may be configured as follows.
The information processing device is equipped with a unit score processing unit that acquires an actual unit score of a target user for at least one unit, acquires a first predicted unit score of the target user for a prediction target unit different from the at least one unit based on the score relationship between units based on the actual unit scores of multiple users for the multiple units and the acquired actual unit score of the target user, acquires actual unit scores of multiple other users different from the target user for the prediction target unit, acquires a second predicted unit score of the target user for the prediction target unit based on the score relationship between users based on the actual unit scores of the multiple users for the multiple units and the acquired actual unit scores of the multiple other users, and acquires a third predicted unit score of the target user for the prediction target unit based on the first predicted unit score and the second predicted unit score.

[Other embodiments]
A unit represents a collection of contents, and is not limited to being included in a subject. A unit may be a subject or a school subject.

Although points have been described as an example of a score, this is not limiting. The score may be a percentage value indicating the level of mastery of the unit.

The information processing device may be realized by one device as described in the above example, or may be realized by multiple devices with distributed functions. Although an example has been described in which the storage device 14 of server 1 stores various data, a server different from server 1 may store various data in place of server 1. Multiple servers different from server 1 may store various data in a distributed manner in place of server 1.

The program may be transferred in a state where it is stored in an electronic device, or in a state where it is not stored in an electronic device. In the latter case, the program may be transferred via a network, or in a state where it is recorded on a recording medium. The recording medium is a non-transitory tangible medium. The recording medium is a computer-readable medium. The recording medium may be in any form, such as a CD-ROM or memory card, as long as it is capable of storing the program and is computer-readable.

In short, this invention is not limited to the present embodiment as it is, and the components can be modified and embodied in the implementation stage without departing from the gist of the invention. In addition, various inventions can be formed by appropriately combining multiple components disclosed in this embodiment. For example, some components may be deleted from all the components shown in this embodiment. Furthermore, components from different embodiments may be appropriately combined.
The following is a summary of the claims as originally filed.
[C1]
a unit score processing unit that acquires a score of an output lesson associated with a unit based on the attendance of the output lesson by the target user, and updates a unit score indicating the ability of the target user in the unit based on the score of the output lesson;
an exam score prediction unit that obtains a predicted exam score for the target user based on the unit scores of a plurality of units for the target user;
An information processing device comprising:
[C2]
The information processing device described in [C1], wherein the unit score processing unit updates the unit score of the unit based on a score adjusted for the output lesson based on the difficulty level set for the output lesson and the contribution of the output lesson to the unit score of the unit.
[C3]
The information processing device according to [C1] or [C2], wherein the unit score processing unit updates the unit scores of the plurality of units of the target user based on the degree of score decrease over time according to the degree of forgetting.
[C4]
The information processing device described in [C3], wherein the unit score processing unit reduces the degree of decrease for the unit based on taking the output lesson.
[C5]
An optimization processing unit optimizes the contribution of the output lessons associated with the plurality of units to the unit scores of the plurality of units, the difficulty of the output lessons associated with the plurality of units, and the forgetting rate by machine learning based on the predicted test scores of the plurality of users and the actual test scores of the plurality of users.
An information processing device according to [C3] which cites [C2].
[C6]
The unit scores of the plurality of units include a unit score based on an actual unit score of the target user for at least one unit and a unit score based on a predicted unit score of the target user for a predicted target unit different from the at least one unit,
The unit score processing unit acquires an actual unit score of the target user, and acquires a predicted unit score of the target user for the prediction target unit based on a relationship between the scores of the units based on the actual unit scores of the multiple users for the multiple units and the acquired actual unit score of the target user.
The information processing device according to [C1].
[C7]
The unit scores of the plurality of units include a unit score based on an actual unit score of the target user for at least one unit and a unit score based on a predicted unit score of the target user for a predicted target unit different from the at least one unit,
the unit score processing unit acquires actual unit scores of a plurality of other users other than the target user for the unit to be predicted, and acquires a predicted unit score of the target user for the unit to be predicted based on a relationship between scores of a plurality of users including the target user and the plurality of other users for the plurality of units based on the actual unit scores of the plurality of users including the target user and the plurality of other users and the acquired actual unit scores of the plurality of other users;
The information processing device according to [C1].
[C8]
obtaining a score for an output lesson associated with a unit based on the target user's attendance of the output lesson;
updating a unit score indicating the target user's ability in the unit based on the score of the output lesson;
obtaining a predicted exam score for the target user based on the unit scores of a plurality of units for the target user;
An information processing method comprising:
[C9]
On the computer,
obtaining a score for an output lesson associated with a unit based on the target user's attendance of the output lesson;
updating a unit score indicating the target user's ability in the unit based on the score of the output lesson;
obtaining a predicted exam score for the target user based on the unit scores of a plurality of units for the target user;
An information processing program that executes the above.

1...server, 2...terminal, 11...processor, 12...ROM, 13...RAM, 14...storage device, 15...communication interface, 21...processor, 22...ROM, 23...RAM, 24...storage device, 25...communication interface, 26...input device, 27...display device, 111...unit score processing unit, 112...exam score prediction unit, 113...output unit, 114...optimization processing unit, 115...exam score prediction model processing unit, 116...unit score prediction model processing unit, 141...management data, 142...scoring ratio data, 143...exam score prediction model, 144...first similarity data, 145...first unit score prediction model, 146...second similarity data, 147...second unit score prediction model.

Claims (6)

a unit score processing unit that acquires a score of an output lesson acquired by a target user based on the target user's attendance of an output lesson associated with a unit, and updates the unit score of the unit by reflecting the acquired score of the output lesson on the unit score of the unit immediately before the attendance of the output lesson using a parameter indicating a degree to which the acquired score of the output lesson is likely to be reflected in the unit score of the unit immediately before the attendance of the output lesson;
a test score prediction unit that obtains a predicted test score for the target user at a time point when the unit scores of the units are updated based on the unit scores of the target user for a plurality of units;
An information processing device comprising: The parameters include a difficulty level,
The information processing device of claim 1 , wherein updating the unit score of the unit includes adjusting the score of the acquired output lesson based on the difficulty level, and updating the unit score of the unit based on the adjusted score . The parameters include a contribution degree,
The information processing device of claim 2, wherein updating the unit score of the unit includes updating the unit score of the unit to a score obtained by combining the adjusted score based on the contribution rate and the unit score of the unit immediately before taking the output lesson . The information processing device of claim 1 , further comprising an optimization processing unit that optimizes the parameters used to update the unit score of the unit so as to reduce the difference between each user's predicted test score and actual test score by machine learning based on the predicted test scores of a plurality of users and the actual test scores of the plurality of users. An information processing method executed by an information processing device,
Obtaining a score of an output lesson acquired by a target user based on the target user's attendance of the output lesson associated with a unit;
updating the unit score of the unit by reflecting the acquired score of the output lesson to the unit score of the unit immediately before taking the output lesson using a parameter indicating a degree to which the acquired score of the output lesson is likely to be reflected in the unit score of the unit immediately before taking the output lesson;
Obtaining a predicted test score for the target user at a time when the unit scores of the units are updated based on the unit scores of the target user for the multiple units;
An information processing method comprising: An information processing program capable of causing a computer to execute processing by each unit included in the information processing device according to any one of claims 1 to 4 .